Anorexia nervosa (AN) and bulimia nervosa (BN) are prevalent disorders in adolescent and young adult women and are associated with high rates of morbidity and mortality. This study represents the only prospective, naturalistic, long-term study with interviews conducted at frequent intervals to map the course and outcome of these disorders. We propose to extend our current eight-year longitudinal study of 250 women with AN and/or BN for up to 13 1/2 years. The aims are: l) to describe the course of each disorder; 2) to ascertain whether AN and BN are distinct phenomena; 3) to assess static and dynamic prognostic factors; 4) to assess the relationship between eating disorders and comorbid disorders; and 5) to empirically determine definitions for recovery, relapse, and remission. In the first five years, we recruited 229 patients meeting DSM-III-R criteria for AN and/or BN and divided them into three diagnostic categories: AN, BN, and AN/BN, completed the schedule of interviews described in the original proposal, assembled master files of intake and three, six, nine, 12, 15, 18, 21, and 24 month follow-up data, analyzed intake, follow-up, and treatment data collected, and conducted one reliability study. Over the past three years, we have maintained our cohort, collected all thirty-six month data, and conducted rudimentary analyses. In order to increase the study power for the AN group, twenty one additional ANs were recruited for a total of 250 subjects. The original sample has been followed between 4 1/2 and seven years and the new ANs have been followed for over three years. The attrition rate is a low 5.7%. Based on DSM-IV subtyping of AN, the sample was reclassified into three groups: AN-Restricting type (ANR), AN-Binge eating/purging type (ANB), and BN. High levels of current comorbid Axis I disorders (64%) and relatively low rates of Axis II disorders (25%) were observed. While comorbidity did not differentiate AN and BN, course of illness did. In the first two years of follow-up, a great majority of subjects improved symptomatically. Rates of partial recovery (less than full DSM-IV criteria for more than or equal to 8 weeks) were: 54% of ANR, 81% of ANB, and 88% of BN subjects. Rates of full recovery (asymptomatic for more than or equal to 8 weeks) were: 8% of ANR, 17% of ANB, and 57% of BN subjects. The data indicate that the diagnosis of AN has serious implications and that BN has a moderate rate of recovery and high rate of relapse (38%). Because of the limited data to support the current diagnostic criteria, and their implications for treatment, it is imperative to follow course, level of continuing impairment, and comorbidity of subsyndromal AN and BN. Continued longitudinal follow-up is essential to maintain this unique cohort, to allow a sufficient number of subjects to recover and relapse, to comprehensively describe course of illness, identify predictors, and to develop a consistent set of definitions for change points in the course of AN and BN. We propose to continue conducting evaluations for an additional five years on all probands. Data gathered from the study will allow us to improve our design of research protocols on pathophysiology and treatment, increase patient and public education about AN and BN, and contribute to the nosology of the eating disorders.